Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
From:   =?utf-8?B?7J207KCV7ZiE?= <jhs2.lee@samsung.com>
To:     "'Sami Tolvanen'" <samitolvanen@google.com>
Cc:     "'Mike Snitzer'" <snitzer@redhat.com>,
        "'Alasdair Kergon'" <agk@redhat.com>,
        "'device-mapper development'" <dm-devel@redhat.com>,
        "'LKML'" <linux-kernel@vger.kernel.org>
In-Reply-To: <CABCJKucMyQbtt-LACCRSk6bxAqmS05eVhv-5ou3_XZ1Nz+XYug@mail.gmail.com>
Subject: RE: [PATCH 2/2] dm verity: allow only one verify mode
Date:   Fri, 12 Mar 2021 16:04:08 +0900
Message-ID: <009001d7170d$e584d360$b08e7a20$@samsung.com>
MIME-Version: 1.0
Content-Transfer-Encoding: quoted-printable
Thread-Index: AQJywKbl46ZRzKPvCQNOoTqt3A/c2AF7PY8mAhqlor8B7jzycakcb2ag
Content-Language: ko
Content-Type: text/plain; charset="utf-8"
CMS-TYPE: 101P
DLP-Filter: Pass
References: <CGME20210311121850epcas1p493c255a586998916febfebaf994bc5dc@epcas1p4.samsung.com>
        <1615464651-23675-1-git-send-email-jhs2.lee@samsung.com>
        <1615464651-23675-2-git-send-email-jhs2.lee@samsung.com>
        <CABCJKucMyQbtt-LACCRSk6bxAqmS05eVhv-5ou3_XZ1Nz+XYug@mail.gmail.com>
Precedence: bulk

Hello, Dear Sami Tolvanen.
Thank you for reply.

> I agree that we shouldn't allow this, at least not without a warning, but
> out of curiosity, do you actually have a situation where this could happe=
n?
> One ideally shouldn't be passing untrusted parameters to dm-verity.

Of course, I don't think this will happen because they are dm-verity expert=
s.
But since we are humans, I think this case could happen accidentally.
So it would be a good at preventing these cases.

> I don't have a strong opinion about this, but the documentation doesn't
> talk about verity modes, so perhaps this could be reworded to something
> like =22Conflicting error handling parameters=22?

Yes of course. That looks better.

I also had some ambiguous about how to express it.=20
This is because I couldn't find it in document.=20
The code says verity mode, so I wrote it down. never mind it :)=20

like this)
    case DM_VERITY_MODE_LOGGING:
    case DM_VERITY_MODE_RESTART:
    case DM_VERITY_MODE_PANIC:


> On Thu, Mar 11, 2021 at 4:19 AM JeongHyeon Lee <jhs2.lee=40samsung.com>
> wrote:
> >
> > If there are multiple verity mode when parsing the verity mode of dm
> > verity table, it will be set as the last one.
> > So set to 'allow only once' to prevent it.
>=20
> I agree that we shouldn't allow this, at least not without a warning, but
> out of curiosity, do you actually have a situation where this could happe=
n?
> One ideally shouldn't be passing untrusted parameters to dm-verity.
>=20
> >
> > Signed-off-by: JeongHyeon Lee <jhs2.lee=40samsung.com>
> > ---
> >  drivers/md/dm-verity-target.c =7C 38
> > ++++++++++++++++++++++++++---------
> >  1 file changed, 28 insertions(+), 10 deletions(-)
> >
> > diff --git a/drivers/md/dm-verity-target.c
> > b/drivers/md/dm-verity-target.c index 808a98ef624c..b76431dc7721
> > 100644
> > --- a/drivers/md/dm-verity-target.c
> > +++ b/drivers/md/dm-verity-target.c
> > =40=40 -893,6 +893,28 =40=40 static int verity_alloc_zero_digest(struct
> dm_verity *v)
> >         return r;
> >  =7D
> >
> > +static inline bool verity_is_verity_mode(const char *arg_name) =7B
> > +       return (=21strcasecmp(arg_name, DM_VERITY_OPT_LOGGING) =7C=7C
> > +               =21strcasecmp(arg_name, DM_VERITY_OPT_RESTART) =7C=7C
> > +               =21strcasecmp(arg_name, DM_VERITY_OPT_PANIC)); =7D
> > +
> > +static int verity_parse_verity_mode(struct dm_verity *v, const char
> > +*arg_name) =7B
> > +       if (v->mode)
> > +               return -EINVAL;
> > +
> > +       if (=21strcasecmp(arg_name, DM_VERITY_OPT_LOGGING))
> > +               v->mode =3D DM_VERITY_MODE_LOGGING;
> > +       else if (=21strcasecmp(arg_name, DM_VERITY_OPT_RESTART))
> > +               v->mode =3D DM_VERITY_MODE_RESTART;
> > +       else if (=21strcasecmp(arg_name, DM_VERITY_OPT_PANIC))
> > +               v->mode =3D DM_VERITY_MODE_PANIC;
> > +
> > +       return 0;
> > +=7D
> > +
> >  static int verity_parse_opt_args(struct dm_arg_set *as, struct
> dm_verity *v,
> >                                  struct dm_verity_sig_opts
> > *verify_args)  =7B =40=40 -916,16 +938,12 =40=40 static int
> > verity_parse_opt_args(struct dm_arg_set *as, struct dm_verity *v,
> >                 arg_name =3D dm_shift_arg(as);
> >                 argc--;
> >
> > -               if (=21strcasecmp(arg_name, DM_VERITY_OPT_LOGGING)) =7B
> > -                       v->mode =3D DM_VERITY_MODE_LOGGING;
> > -                       continue;
> > -
> > -               =7D else if (=21strcasecmp(arg_name, DM_VERITY_OPT_REST=
ART)) =7B
> > -                       v->mode =3D DM_VERITY_MODE_RESTART;
> > -                       continue;
> > -
> > -               =7D else if (=21strcasecmp(arg_name, DM_VERITY_OPT_PANI=
C)) =7B
> > -                       v->mode =3D DM_VERITY_MODE_PANIC;
> > +               if (verity_is_verity_mode(arg_name)) =7B
> > +                       r =3D verity_parse_verity_mode(v, arg_name);
> > +                       if (r) =7B
> > +                               ti->error =3D =22Already verity mode se=
t=22;
>=20
> I don't have a strong opinion about this, but the documentation doesn't
> talk about verity modes, so perhaps this could be reworded to something
> like =22Conflicting error handling parameters=22?
>=20
> > +                               return r;
> > +                       =7D
> >                         continue;
> >
> >                 =7D else if (=21strcasecmp(arg_name,
> > DM_VERITY_OPT_IGN_ZEROES)) =7B
> > --
> > 2.17.1
> >
>=20
> Sami